Website Redesign SEO: How to Relaunch Without Losing Rankings

You keep your Google rankings through a redesign by keeping the things Google ranks. Leave URLs alone wherever you can and give every one that changes a one-to-one 301 redirect. Keep the substance of the pages that already bring in traffic. The new design is rarely what costs you. Lost rankings almost always trace back to something mechanical, usually a URL that vanished without a redirect or a staging setting that shipped to the live site.

Even a careful launch wobbles for a while. Google's own guide to moving a site tells owners to expect temporary ranking fluctuation while it recrawls and reindexes, and says a medium-sized site can take a few weeks or more before Google starts showing the new URLs instead of the old ones. A dip that recovers inside that window is normal. A drop that keeps sinking, or lands hard on particular pages, has a cause. Most of this guide is about making sure you never have to go looking for it.

Why redesigns lose rankings

Google ranks URLs, plus the content and links attached to each one, and a redesign touches all of it at once. These are the failures we find most often when a business calls us after a launch went wrong.

• URLs changed without redirects. The new CMS turns /services.html into /services/, nobody maps the old address, and its links and rankings now hit a 404. Google's documentation on redirects recommends a permanent server-side redirect whenever a URL changes for good.
• Content removed or thinned out. A 1,200-word service page that ranked for dozens of queries gets cut to three lines and an icon grid because it looked cleaner in the mockup. The rankings leave with the words.
• Broken internal links. Links inside old posts still point at old URLs, or the new menu uses script-driven buttons instead of real links. Google's link best practices say it can generally only crawl links that are <a> elements with an href attribute.
• Staging blocks left in place. The noindex tag or robots.txt rule that kept the development site private ships to production on launch day. Nothing drops a whole site out of Google faster.
• Content that only appears after JavaScript runs. Google renders JavaScript, but its JavaScript SEO guide says pages can sit in a rendering queue for longer than a few seconds, and still recommends server-side or pre-rendering because not all bots can run JavaScript. A site that ships an empty HTML shell is betting its rankings on that queue.
• Rewritten titles and headings. A title that ranked for "emergency plumber Denver" for six years gets swapped for a brand slogan. Google builds its title links from sources that include the title element and the page's headings, so changing both at once changes what the page appears to be about.
• Structured data that did not make the trip. Review, product and local business markup often lived in a plugin the new build drops. Google's introduction to structured data warns that markup can break after deployment because of templating or serving issues, and says to test it during development and after launch.

Behind many of these sits a bigger mistake: changing everything at once. Google's site move guide says to change one thing at a time, so a new domain, a new CMS and a new layout should not land on the same day. The Change of Address help page adds that pairing a move with a redesign of content and URL structure will probably cost some traffic while Google relearns the pages. If your redesign does not need new URLs or a new domain, do not give it new ones.

Visitors judge the new design. Google judges the URLs and the words underneath it, and those are the parts you have to carry across.

Inventory everything before you change anything

You cannot protect rankings you have not written down. Before a designer opens a file, build a record of the current site: every URL, what it ranks for, what links to it, and what it earns. That inventory feeds the redirect map, and it becomes the benchmark you judge the launch against.

1. Crawl the live site. Run a crawler such as Screaming Frog and export every URL with its status code, title, H1, meta description, canonical and word count. You will crawl again after launch and compare.
2. Pull URLs from every other source. A crawler only finds linked pages. Google's site move guide suggests adding URLs from your sitemaps, CMS, analytics and server logs, plus the images, PDFs and other files you host. Orphaned pages that still get search traffic turn up here and nowhere else.
3. Export your Search Console data. In the Performance report, set the date range to the last 16 months, the view Google's guide to debugging traffic drops recommends, and export pages and queries. Every page that earns clicks goes on the protect-first list.
4. List the pages other sites link to. Search Console's Links report shows which pages have external links. A linked page with no redirect throws away links that may have taken years to earn.
5. Record landing pages that convert. A quiet page that produces half your inquiries matters more than a popular old blog post. Your analytics will tell you which is which.
6. Snapshot your top pages. Save the title, H1, main copy and structured data of each. If the new version of a page will be shorter, make that call on purpose, with the ranking data in front of you.

Do this at kickoff, not in launch week. Most marketing sites take four to eight weeks from kickoff to launch, and finding out in the final week that 300 blog posts have dated URLs leaves no time to handle them properly.

Build a one-to-one 301 redirect map

The redirect map is a spreadsheet: old URL on the left, new URL on the right, one row for every URL that changes. Google recommends server-side permanent redirects (301 or 308), which tell it the new URL is the one to show in results. A temporary 302 keeps the old URL in results instead.

• One old URL, one closest match. The old kitchen remodeling page goes to the new kitchen remodeling page, not to a general services page.
• Never dump everything on the homepage. Google's site move guide says not to redirect many old URLs to one irrelevant destination such as the home page, because it can confuse users and might be treated as a soft 404. Search Console lists soft 404s among the pages it does not index, so that redirect carries nothing across.
• Merged content is the exception. If three thin pages become one strong page, redirect all three to it. Google explicitly allows this.
• Let dead pages go honestly. A discontinued service or an expired promotion with no real equivalent should return a 404 or 410, which is what Google's guide asks for with deleted content.
• Point straight at the final URL. Googlebot can follow up to 10 hops in a chain, but Google advises redirecting to the final destination directly, and otherwise keeping chains to ideally no more than 3 and fewer than 5. Redesigns are where chains are born: a 2019 redirect to /contact-us/ meets a new one to /contact/. Rewrite the old rules so each lands on the final page in one hop.
• Include files as well as pages. PDFs, images and downloads earn traffic and links too, and Google's guide names them specifically.
• Keep redirects for at least a year. Google says to keep them as long as possible, generally at least one year, so it can transfer all signals to the new URLs, and suggests keeping them indefinitely for visitors' sake.

Here is a slice of a typical map for a renovation company moving off an old CMS:

Test the map before launch. Load the old URL list into your crawler, point it at staging with redirects switched on, and confirm that every row returns a single 301 to a page that returns 200, or the 404 or 410 you planned.

Keep staging out of Google, and out of production

Your staging copy has to stay out of search results, and the usual ways of hiding it each come with a catch. Google's introduction to robots.txt says it "is not a mechanism for keeping a web page out of Google," because a blocked URL can still be indexed if other pages link to it. A noindex tag works, but Google's noindex documentation points out that crawlers have to reach the page to see the tag, so combining noindex with a robots.txt block cancels it out.

Put staging behind a password instead, with HTTP authentication or your host's built-in protection. Google lists password protection next to noindex as a way to keep pages out of its results. And because the password lives in the server setup rather than your templates, it does not travel to the live site with the code.

Tags and rules inside the code do travel, and that is the real risk. We check every launch for these four:

• A noindex robots meta tag on every page. On WordPress this is often the "Discourage search engines from indexing this site" checkbox, which the WordPress documentation says adds a noindex, nofollow robots meta tag.
• A robots.txt rule of Disallow: / that blocks crawling of the entire site.
• An X-Robots-Tag: noindex header set at the server or CDN. It never appears in the page source, which is why it gets missed.
• Canonical tags or sitemaps that still point at the staging domain.

Leftover noindex and robots.txt blocks sit first on the list of common mistakes in Google's site move guide.

Your launch-day checklist, phase by phase

Most of the work happens before launch, so launch day itself should be boring. If your traffic is seasonal or dips on certain weekdays, Google suggests moving during the quiet period, when fewer visitors will hit early problems and more server capacity is free for Googlebot.

Two rules for the day itself. Redirects go live in the same release as the new site, never "next week." And someone checks the live site for staging blocks within the first hour, before anyone celebrates.

What to watch in the weeks after launch

Google's guide says visibility may fluctuate during a move and that rankings settle down over time, with medium-sized sites taking a few weeks or more and larger sites longer. Check daily for the first two weeks, then weekly until your top pages are back at their baseline.

The Page indexing report

Search Console's Page indexing report is your early warning system. "Page with redirect" next to your old URLs is good news: Google has seen the redirect. These statuses need action:

• Not found (404). A spike means old URLs are missing from the redirect map, or links point at pages that no longer exist. Google says it frequently sees redirects pointing at wrong, non-existent URLs.
• Soft 404. Often homepage or category redirects that Google did not accept as real equivalents, or new pages so thin they look empty.
• URL marked 'noindex' or blocked by robots.txt. If pages you want ranked show up here, a staging setting shipped. Fix it the same day.
• Redirect error. Usually a loop, or a chain that runs too long.

Sitemaps, crawling and traffic

Google's guide describes a handy trick: keep a sitemap of the old URLs submitted next to the new one for a while, and watch its indexed count fall toward zero as the new count climbs. Check server logs for Googlebot hitting error pages, and make sure your hosting has headroom, because Google temporarily crawls a migrated site more heavily than usual. In the Performance report, compare clicks since launch with the same period before, page by page.

Change of Address is for domain moves only

If you keep your domain, you do not need the Change of Address tool. Google says it is for moving from one domain or subdomain to another, and not for HTTP to HTTPS moves, www to non-www changes, or moving pages within the same domain. For a same-domain redesign, redirects and an updated sitemap do the job.

When a drop does persist, line the losing pages up against your inventory. On the sites we audit, the cause is nearly always on the list at the top of this post, most often a missing redirect or a page that lost its content.

Use the redesign to fix what was holding you back

A redesign is also the cheapest moment to fix structural problems, because the templates are being rebuilt anyway. Two improvements pay off most often.

Speed. New templates are a chance to ship lighter pages, with right-sized images and far less JavaScript. Our post on why website speed sells covers the revenue math and a fix list in order of return. If the redesign also means leaving a page builder, our comparison of custom WordPress and page builders explains what carries over and what gets rebuilt.

Content structure. Keep the content that ranks, then make it easier to use: answer the question near the top of each page and write headings that match what customers actually type. That structure also helps you get quoted in AI answers, which we cover in SEO vs AEO vs GEO.

Our rule is to improve pages rather than replace them. Expand a thin page, tighten a rambling one, but keep its URL and its topic. For the same reason, and because Google advises changing one thing at a time, we push back when a redesign brief includes a domain change. Launch the redesign, let rankings settle, then move the domain as a separate project if you still want to.

Start with the map, not the mockups

If a redesign is on your calendar, start the inventory this week, before anyone debates colors. It takes a small slice of the project and protects the part of your site that already pays for itself: the pages people find on Google today.

Every website we build as a redesign starts from that inventory and redirect map, and the before-and-after checks are part of how we run SEO for our clients. Once the new site has settled, our technical SEO checklist picks up where this one stops. The aim is a launch week where nothing happens in Search Console worth talking about.